BCG’s AI strategy is decentralized across multiple organizations, with no single group directing the entire agenda.
OpenAI remains BCG’s primary model provider, while Anthropic began emerging as a notable second option in Q1 2026.
Although BCG relies primarily on external AI vendors, one key focus of its internal AI development is knowledge-management solutions and supporting infrastructure that leverage the firm’s expertise.
BCG has largely avoided training its own models, focusing its model-training capabilities on client work instead.
BCG has invested substantially across a wide range of AI initiatives without appointing a Chief AI Officer, a role some companies have introduced. Responsibility is instead distributed across the firm rather than concentrated in a single entity.
The Tech Build and Design Division of BCG. Launched in 2022, the unit develops AI products for clients.
BCG's global center for the design and build of AI platforms. It helps clients choose a platform, design the solution, and build it.
BCG X's scientific research arm, working with universities, companies, and governments to speed up scientific discovery.
The central engineering group that builds and maintains the AI tools BCG's practice areas run on.
Oversees the AI transformation of BCG's industry practices, including turning expertise into sellable products.
Primary team that drives all intellectual property development and manages go-to-market for AI globally across the company.
Partners with BCG's internal corporate functions, including People, Finance, Legal, and Risk, to design and deliver AI solutions.
An engineering group assembled across BCG, BCG X, and the technology practice to build the firm's internal AI platform.
Independently reviews the AI solutions that other BCG teams build, judging them on fairness, safety, and reliability.
A global network of trained employees, drawn from every department, who help colleagues use AI in their daily work.
A required certification for all 30,000 BCG employees, running from basic AI literacy up to building working AI systems.
A firm-wide internal competition that surfaced the best ways employees had found to use AI in their own work.
Although the diagram captures only part of BCG’s expansive AI footprint, the practices organization appears to be consolidating AI development. To reduce duplicate system development across practice groups, BCG quietly formed the AI Factory, a centralized team tasked with building and managing shared AI capabilities across the practices. Furthermore, Dylan Bolden expanded his scope in May 2026 to oversee Applied AI, in addition to his existing role as chair of functional practices.
BCG maintains a fairly streamlined AI infrastructure. Rather than relying on a fragmented assortment of tools across different divisions, the firm typically uses a single standardized tool for each key AI function.
Although BCG operates in a multi-cloud environment, AWS provides a great deal of its AI technology. Applications connect to models through Amazon Bedrock, data is ingested via AWS Glue, and employees often leverage Amazon OpenSearch to query internal firm documents.
OpenAI remains the incumbent provider at the model layer. Following a 2023 enterprise agreement that deployed ChatGPT to all its employees, BCG staff have created approximately 36,000 custom GPTs, primarily built by consultants rather than a centralized team.
In Q1 2026, Anthropic began to establish itself as the clear second-place provider. Expectations of Claude proficiency became more widespread among BCG's non-technical staff, while BCG's engineering teams began treating Claude Code as a standard build tool.
BCG internal stack: Technologies BCG teams use to build AI solutions for internal use.
Client delivery stack: Technologies BCG expects employees to know when building AI systems for clients.
BCG’s client work extends into real-time data processing and custom model training, capabilities that appear to play a limited role in BCG's own operations.
BCG’s AI strategy centers on buying established solutions and only building internally where it sees clear differentiation. Unlike many large corporations that have invested heavily in proprietary AI models, BCG has placed little emphasis on training models for its own use. A similar pattern is evident at Bain and McKinsey, with all three firms relying primarily on established providers such as OpenAI, Anthropic, and Google.
BCG has prioritized its build decisions around leveraging the company's internal knowledge. BCG built a slide creation tool called Deckster, which draws on 800+ BCG templates to enable consultants to build presentations. BCG’s first chatbot, GENE, was built on GPT-4o and customized with proprietary interviews and internal research.
In addition to those internal tools, BCG made retrieval-augmented generation (RAG) a focus area. This technology lets BCG put its institutional knowledge to work, and a consultancy's expertise is often considered one of its most valuable assets. In October 2024, the firm stood up a research group whose founding job was building knowledge graphs from its slide decks and meeting transcripts. Neo4j sits on the knowledge-graph side, and LlamaIndex on the document side.
BCG has taken a broad approach to partnerships in the AI space, particularly with leading frontier model providers, maintaining substantial relationships with both OpenAI and Anthropic simultaneously. This is unlike McKinsey and Bain, which have concentrated primarily on OpenAI. With OpenAI, BCG holds Elite Partner status (the top tier of OpenAI's partner network) and is part of the firm’s Frontier Alliance. With Anthropic, BCG has committed $500 million by 2030 to AI initiatives focused on social impact. The firm also has a separate Claude-integrated cybersecurity collaboration. BCG also announced an expansion of its partnership with Google Cloud in April 2026, under which Google embeds its engineers within BCG's teams and provides BCG with early access to its frontier models.
However, BCG didn't follow Bain and McKinsey into investing in OpenAI's Deployment Company, a $4 billion-plus venture that embeds OpenAI engineers directly into client operations to deploy AI in production.